Robotics as a Learning Medium for Engineering Practice
and Team-based Design in Capstone Projects
Ninad Pradhan, Timothy Burg, Richard Groff, Darren Dawson
Dept. of Electrical and Computer Engineering, Clemson University
In the capstone project courses in Electrical and Computer Engineering at Clemson University robotics projects are
used as a medium to teach senior-year undergraduates the tools required for engineering practice and team-based
design, qualities essential for their professional success as engineers. The student teams solve technical problems
while learning about system design, ethics, safety, hardware and software integration, and technical documentation.
The inherent breadth of robotics allows us to control the parameters and skill sets required to execute a project. This
model makes it possible to impart specific skills each semester without having to make drastic curricular changes. In
addition, evaluation, and feedback from industry members serves to measure student performance from a second
viewpoint. This paper documents our experience of using robotics as a learning medium for senior design.
Corresponding Author: Timothy Burg, tburg@clemson.edu
Introduction
Interdisciplinary skills are necessary for a successful
engineer in today’s industrial environment, and such
skills need to be taught to engineering students as an
integral part of their training. For over twenty years
robotics design projects have been used in Clemson
University’s Electrical and Computer Engineering
(ECE) capstone design classes as the means to connect
classroom learning with practical applications. In this
paper, we outline our approach towards satisfying the
objectives of teaching the skills and tools of modern
engineering practice and team-based design through the
medium of robotics. These projects benefit from active
industry involvement in evaluating both the design
process and the project prototypes.
Representatives from local industry have been
involved with our capstone projects in consistent and
diverse ways. Industry collaborators i) support
evaluation of students on a basic skill set; ii) provide
credibility to expectations on professional presentations,
skills, etc.; and iii) give presentations on sales,
patenting, etc. The close involvement of industry
representatives not only provides vital feedback to
students regarding the merit of their design, but also
provides the instructors inputs on the development of
future project ideas.
From a design and control perspective, robotics can
be classified as one of the many sub-problems in
mechatronics, which has been steadily gaining in
importance both as an academic discipline and as a
philosophy for engineering education [1] [2]. The wide
range of engineering abilities encompassed by robotics
makes it an interesting platform for a capstone design
class. The versatility and interdisciplinarity of robotics
makes it possible for us to manipulate the requirements
of projects such that different sets of skills can be
developed over the course of the senior year. This idea
of targeted learning using the medium of robotics is the
central contribution of our approach.
Capstone projects [3] provide the prospective engineer
an opportunity to learn about the design process. The
most favored model for teaching design is currently the
paradigm of Project Based Learning (PBL). Dym [4]
proposes a definition for design which places emphasis
on the technical details of a project as well as the
objective of satisfying a client’s requirements. With this
in mind, our design course gives considerable
importance to the project life cycle and sets specific
documentation and process requirements for teams to
meet during the semester.
Senior design projects in Clemson ECE are spread
over two semesters, with students working on a
different problem in robotics each semester. To
substantiate the process outlined above, we present a set
of example first semester design projects from the
program, all of which are designed to support
assessment of multiple ABET outcomes [5].
Course Philosophy and Rationale
ABET, the accrediting agency for applied science,
computing, engineering, and technology programs [5],
defines criteria that ensure “programs produce graduates
who are ready to enter their professions.” Such criteria
serve as a good reference point for evaluating the
success of our capstone courses. Based on project and
individual scores and feedback from external evaluators,
students are rated on performance on eleven of the
ABET criteria.
To achieve these objectives, Senior Design comprises
two two-credit hour courses: ECE 495 and ECE 496.
Students implement one project each semester in their
senior year. The two projects are not connected by a
continuity of the problem to be solved; rather, they are
connected by the increased level of expertise required
for successful project solution. Generally speaking, the
first semester projects are “static” and the second
semester projects are “dynamic”. Additionally, second
semester projects require a top-down design approach as
no components are specified a priori while the first
semester designs are based on a few given parts,
typically a motor and camera. This transitions the teams
from previous labs where all components are given and
a bottom-up design approach is executed.
Robotic projects specified by the instructor are used
to delimit the solution space and to direct training in the
tools required for such solutions. Specifically, students
must
gain
expertise
in
such
tools
as
MATLAB/Simulink,
C++
programming,
data
acquisition, and image processing. The projects hinge
on a demonstrable proficiency in technical skills that
include closed-loop motor control, real-time control
software, electronic circuit design, and systematic
design approach.
Students are arranged in teams of five students in
order to practice generating and implementing technical
solutions in a collaborative manner. Each student
completes the Comprehensive Assessment for TeamMember Effectiveness (CATME) survey for their team
members three times during the semester [6]. The
students complete the design process steps described in
[7]
. Highlights include the use of the BEES software
published by NIST [8], to make a decision regarding
material for the project shipping container; use of the
Design Failure Modes Analysis (DFMEA) to help
analyze the safety of their designs; creation of a Gantt
chart to help organize their activities. Each step in the
design process is documented as a written report; all of
which are then incorporated into a group website at the
end of the semester.
In addition to the common learning elements
described above, project specifications are defined each
semester such that one specific skill from this set is
highlighted. This is consistent with the paradigm of
Project Based Learning [4], the objective of which is to
define a problem such that the process of solving it
leads to acquisition of certain skills. In the following
section, we outline some of the tenets and skills of
engineering practice and design, and use selected
projects from ECE 495 in previous semesters [9] to
explain how robotics plays a central role in educating
students in these areas.
Engineering Practice
Systems Integration – The Laparoscopy Surgery
Robot
The difference between conventional system design and
mechatronics design is that the latter requires, as
outlined in Isermann’s work [10], “simultaneous
engineering” to take place. That is, the project must be
considered in its totality right from the beginning, rather
than as a collective of components designed by different
kinds of engineers. Thus, the idea of systems integration
can hardly be separated from the notion of a robotics or
mechatronics system design.
To emphasize the multidisciplinary practices
associated with systems integration, students were asked
to design a low-cost joystick-controlled laparoscopic
surgery leader-follower robot. The motivating idea for
this project came from the widespread use of
laparoscopic surgery tools in medicine, and the rapidly
growing usage of teleoperated laparoscopy robots.
To make this project work, groups had to
simultaneously address the challenges of adhering to
client specifications for the user interface, interfacing
multiple motor-encoder pairs, developing software to
work smoothly with the electronics, and design a
mechanical master device which was safe to handle and
functionally complete.
Customer Requirements – Haptic Virtual
Manipulatives
Interaction with a real customer is beneficial for a senior
design class in two ways – it provides the students a
chance to deliver technology which potentially satisfies
real social or industrial requirements and it gives them a
chance to practice their design and implementation
constrained by strict requirements imposed by the
customer.
The designer of Math Out of the Box® [11], a K–5,
inquiry-based math curriculum, served as the project
customer for this system. The task assigned to the
groups was to design haptic virtual manipulatives for K12 students to learn about symmetry and equivalence
(heavy vs. light, for example). As a result of low student
performance in K-12 mathematics, experts recommend
the development of new tools for teaching such topics
as number sense, fractions, algebra, and geometry and
measurement. One of the most efficient methods
currently available for such instruction is the use of
manipulatives [12], using real-world objects to teach
abstract concepts. Haptic virtual manipulatives [13] are
computer generated objects or shapes whose
interactions can be felt by users through a simple haptic
device.
Groups were provided with an OpenGL interface
template which displayed either the Symmetry or the
Balance virtual environment. The hardware to be
designed was a simple haptic interface to allow tactile
interaction with the virtual world seen on the monitor by
delivering a specific torque profile. Additionally, it was
necessary to have enhanced safety features for the
device, since it was to be handled by young
schoolchildren.
A panel of practicing engineers
evaluated the final designs. For this project, the class
was awarded a 2010 NCEES (National Council of
Examiners for Engineering and Surveying) Engineering
Award for Connecting Professional Practice and
Education [14].
with one or more PID-controlled degrees-of-freedom
serves to strongly reinforce their conceptual
understanding of control systems. One such practical
system is the puzzle solving robot, in which a singleDOF rotary robotic arm with quadrature encoder
feedback is controlled.
Earlier in the semester, teams were asked to
implement a PID-controller using encoder feedback to
mimic a clock-hand with a variable, user-defined stepsize. This provided experience with understanding how
to implement a closed-loop PID control MATLAB
Simulink model, and to tune the individual gains to
obtain the desired step response. Further, students used
this project to understand how to interpret the quality of
their controller using plots and graphs in addition to the
visual quality of the response.
Computer Vision – Ping Tac Toe
In industrial systems, the use of cost-effective yet
powerful sensors is a pivotal requirement for
commercial viability of a product. Computer vision,
which uses the camera as a low-cost sensor, fulfills this
requirement. In this project, computer vision was used
as the sensing system for robots capable of
autonomously playing a game of tic-tac-toe with pingpong balls of two different colors.
Groups are introduced to image processing by way of
in-class tutorials covering image processing concepts
from elementary to intermediate levels of complexity,
examples include color spaces, image storage, pixel
access (all elementary), PCA-based region description,
background subtraction, morphological operations, edge
detection (all intermediate). Both MATLAB, which has
an extensive image processing toolbox, and OpenCV,
which is a widely used C++ library for image
processing, are introduced as options for the
implementation software.
The robotic system had to then use the camera input
as raw sensor data and process it using image
processing techniques to identify the game state. The
game state would then be used to plan and execute the
next move in accordance with the game strategy. An
additional element of complexity was introduced by
placing lights at the side of the board for the system to
automatically identify which robot was supposed to
make a move. The game had to be played by completely
autonomous robotic systems which would be able to
identify a winning configuration and stop shooting.
Closed-Loop Control – Puzzle Solving Robot
Closed-loop control is at the heart of all senior design
projects at Clemson ECE. Students are introduced to the
theory of linear control via a senior level course, ECE
409. However, implementing a practical robotic system
This intermediate mini-project was useful training for
their final project, which was the design and
implementation of a puzzle solving robotic arm using
computer-vision as the sensor. The robot was required
to move washers on a ring-shaped game board from
their start position to a desired end position inside
hollow wells located at known locations around the
ring. Strict restrictions were placed on what constituted
successful placement of washers (partially inside the
ring counted as failure) to place greater emphasis on the
performance of their closed-loop controller in this
project.
Discussion
An ongoing challenge is to fully determine whether
students learned what we expected. Many of the
technical and nontechnical aspects are evaluated
through the course grading; however, the real difficulty
lies in determining if we have reshaped the student’s
attitude about their imminent future as a practicing
engineer. In particular, have they taken two important
steps: i) realized that they must draw on a variety of
sources and experiences to solve real problems (i.e. the
solution is not in the preceding chapter of the textbook)
and ii) gained confidence in applying what they have
learned to this point in their undergraduate program out
of the context of the course in which they learned it. To
this end, we administer an anonymous, on-line survey
through Blackboard at the end of each semester using
questions proposed by the University to assess
instructor effectiveness and by the ECE Department and
the Instructor to assess learning outcomes.
Results from the Fall 2011 course survey suggest
some success in our approach to the class. Specific
student responses to the essay question where students
are asked to describe the strengths of the course include
“The course allows lots of freedom in how the projects
are done so you actually have to think about what you
are doing” and “… it was nice to be turned loose and
told to work on projects rather that regurgitate formulas
on examinations”. This is captured quantitatively in the
response to the scaled agreement item (1= “Not at all”
through 5= “Very much”) question that the course
emphasized “Applying knowledge to solve real-world
or realistic problems” which received a 4.6 rating
compared to 4.1 for all students in all classes in the
same major. The question that the course emphasized
“Exercising my creativity in the discipline” received a
4.6 rating compared to 3.5 for all students in all classes
in the same major. Instructor supplied questions, which
don’t have a comparison to other course responses,
include a 4.7 response to the question that “The course
emphasized that teamwork skills are important to an
engineer and the class/lab created an opportunity to
solve a challenging problem as part of a team.” and a
4.7 to the question “The course emphasized that
professional development and continuous learning are
important to an engineering career.”
The survey tool suggests a possible weakness in our
implementation of the approach in that the response to
the question “The judges from IEEE helped reinforce
that the design project has relevance to the real-world”
received only a 4. This response is somewhat surprising
as anecdotally the response of the students to comments
and criticism from practicing engineers is very reverent.
This seems to point to the need for more involvement of
the jury members with the students during the semester.
The students are often skeptical about course elements
such as safety, ethics, and life-cycle and so these topics
can be targeted as an opportunity to increase interaction
and increase the credibility of these elements.
One class of responses that is hardest to interpret is
that the course is “disorganized”, even though lectures,
schedule, and detailed grading criteria are posted at the
start of the course. This criticism may originate from
frustration at solving an open-ended problem with the
inherent uncertainties of our customer-vendor
interaction model; however, it merits investigation. We
are always walking a thin line between over specifying
the solution and creating an intractable problem for the
given time and resources.
Conclusions
We have used robotics in the first semester design
course in order to dictate technical and nontechnical
learning outcomes and used collaboration with industry
partners to evaluate the results. We believe this targeted
learning approach can provide consistency between
projects and team experiences while providing the
design freedom expected in a capstone design course.
We continue to work towards quantifying the outcomes
of this approach as feedback to improving the course.
References
[1] M. Acar and R. Parkin, "Engineering education for
mechatronics," Industrial Electronics, IEEE Transactions
on, vol. 43, no. 1, pp. 106-112, February 1996.
[2] R. Herzallah, "Practically oriented design projects in
mechatronics engineering: a case study of design
experiences and outcomes," International Journal of
Electrical Engineering Education, vol. 47, no. 1, pp. 3146, January 2010.
[3] A. Dutson, R. Todd, S. Magelby and C. Sorensen, "A
review of literature on teaching engineering design through
project-oriented capstone courses," Journal of Engineering
Education, vol. 86, pp. 17-28, 1997.
[4] C. L. Dym, A. M. Agogino, E. Ozgur, D. D. Frey and L. J.
Leifer, "Engineering Design Thinking, Teaching, and
Learning," Journal of Engineering Education, vol. 94, no.
1, pp. 103-120, 2005.
[5] ABET. [Online]. Available: http://www.abet.org/.
[6] CATME. [Online]. Available:
https://engineering.purdue.edu/CATME.
[7] R. M. Ford and C. S. Coulston, Design for Electrical and
Computer Engineers: Theory, Concepts, and Practice,
McGraw Hill Higher Education, 2008.
[8] BEES. [Online]. Available:
http://www.nist.gov/el/economics/BEESSoftware.cfm.
[9] T. Burg. [Online]. Available:
http://www.clemson.edu/ces/crb/ece495/.
[10] R. Isermann, "Modeling and design methodology for
mechatronic systems," IEEE/ASME Transactions on
Mechatronics, vol. 1, no. 1, pp. 16-28, March 1996.
[11] "Math Out of the Box®," [Online]. Available:
http://www.carolinacurriculum.com/math+out+of+the+box
/.
[12] P. Moyer, "What are virtual manipulatives?," Teaching
Children Mathematics, vol. 8, no. 6, p. 372, 2002.
[13] R. B. Singapogu and T. C. Burg, "Haptic virtual
manipulatives for enhancing K-12 special education," in
Proceedings of the 47th Annual ACM Southeast Regional
Conference, 2009.
[14] NCEES. [Online]. Available:
http://www.ncees.org/Licensure/Engineering_Award/Past_
Awards/2010_award_winners.php.